DOCSLIB.ORG

Ontology: Tool for Broad Spectrum Knowledge Integration Barry Smith

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Ontology: Tool for Broad Spectrum Knowledge Integration Barry Smith Foreword to Chinese Translation Ontology: Tool for Broad Spectrum Knowledge Integration Barry Smith BFO: The Beginnings This book was first published in 2015. Its primary target audience was bio- and biomedical informaticians, reflecting the ways in which ontologies had become an established part of the toolset of bio- and biomedical informatics since (roughly) the completion of the Human Genome Project (HGP). As is well known, the success of HGP led to the transformation of biological and clinical sciences into information-driven disciplines and spawned a whole series of new disciplines with names like ‘proteomics’, ‘connectomics’ and ‘toxiocopharmacogenomics’. It was of course not only the human genome that was made available for research but also the genomes of other ‘model organisms’, such as mouse or fly. The remarkable similarities between these genomes and the human genome made it possible to carry out experiments on model organisms and use the results to draw conclusions relevant to our understanding of human health and disease. To bring this about, however, it was necessary to create a controlled vocabulary that could be used for describing salient features of model organisms in a species-neutral way, and to use the terms of this vocabulary to tag the sequence data for all salient organisms. It was with the purpose of creating such a vocabulary that the Gene Ontology (GO) was born in a Montreal hotel bar in 1998.1 Since then the GO has served as mediator between the new genomic data on the one hand, which is accessible only with the aid of computers, and what we might think of as the ‘old biology data’ captured using natural language by means of terms such as ‘cell division’ or ‘mitochondrion’ or ‘protein binding’. Very soon the success, and the utility, of the GO led to the creation of other ontologies covering domains not covered by the GO, such as cell types, proteins, anatomy and diseases, grouped together under the heading ‘OBO’ for ‘Open Biomedical Ontologies’. Beginning in around 2003, however, the problem arose of determining what should serve as the common formal architecture that would bind these ontologies together, and it was in this context that Basic Formal Ontology was born. The result, which is documented at length in the pages that follow, is the Open Biological and Biomedical Ontologies (OBO) Foundry, a suite of life science ontologies all of which descend from BFO and rest on the same set of principles of good practice for ontology development, principles designed to ensure interoperability. The OBO 1 https://www.ncbi.nlm.nih.gov/pubmed/10802651 Foundry ontologies2, like BFO, are created and maintained manually. It is human beings with corresponding scientific and ontological expertise who are responsible for populating the ontologies, formulating definitions and responding to requests from users of the ontology who identify errors or gaps. The underlying idea is that ontologies should be developed in a process of coordinated evolution, each ontology being used in a variety of other co-developed ontologies, for instance through use of its terms in writing definitions. The OBO Foundry principles for coordinated evolution have brought considerable success. Where many ontology initiatives, because they are too closely tied to specific local projects and research groups, enjoy short lives, the aggressive reuse of OBO Foundry ontologies to highly diverse projects and use cases has led to an increasing influence of the OBO Foundry and with this came also the increasing influence of BFO, illustrated not least in the publication of this book. BFO in China As this volume demonstrates, BFO and the OBO Foundry have been discovered, too, in China, where BFO is already being used as top-level ontology by ontologies such as the Traditional Chinese Drug Ontology (TCDO)3 and Chinese Plant Species Diversity Domain Ontology.4 The newly initiated OntoChina program5 aims to support collaborative ontology development, application and distribution across the life sciences and beyond. It is members of the OntoChina group who are responsible for this translation, for the translation into Chinese not only of the BFO ontology itself, but also of other BFO-based ontologies, such as the Information Artifact Ontology (IAO)6 and the Ontology for Biomedical Investigations7. OntoChina also plans to train future ontologists in China using BFO and the BFO-based ontology ecosystem as foundation. BFO in the OBO Foundry It was not until 2016 that BFO was adopted as the official top-level ontology of the OBO Foundry and it is worth examining in this light one core principle of the Foundry suite, namely the principle of orthogonality. Two ontologies are ‘orthogonal’ in Foundry terminology, if they do not overlap – which means: if they share no terms in common. This principle was adopted in support of the idea that there should be exactly one ontology for each domain – so that each term should belong to exactly one ontology. The rationale for this principle is that if all those working in a given domain use the same 2 Barry Smith, et al., “The OBO Foundry: Coordinated Evolution of Ontologies to Support Biomedical Data Integration”, Nature Biotechnology, 25 (11), 2007. 3 Yan Zhu, Lihong Liu, Bo Gao, Yongqun He. TCDO: a community-based Traditional Chinese Drug Ontology. The 11th International Biocuration Conference (Biocuration-2018), Crowne Plaza Hotel Shanghai Fudan, Shanghai, China, April 8-11, 2018. 4 http://manu44.magtech.com.cn/Jwk_infotech_wk3/CN/abstract/abstract4189.shtml 5 http://ontochina.org 6 Werner Ceusters and Barry Smith, “Aboutness: Towards Foundations for the Information Artifact Ontology”, Proceedings of the Sixth International Conference on Biomedical Ontology (ICBO), (CEUR 1515), 2015, 1-5. 7 Anita Bandrowski, Ryan Brinkman, Mathias Brochhausen,et al. “The Ontology for Biomedical Investigations”, PLoS ONE, 11(4): e0154556, April 29, 2016. ontology, then the annotations that they create when they tag their data with ontology terms, will be consistent with each other, and thus cumulate in an orderly fashion. Such cumulativity would allow also the identification of conflicts between different sets of annotations. (Compare the way in which the SI System of Units allows not only the cumulation but also the comparison of quantitative data.) If, however, another core principle of the OBO Foundry requires the use of BFO as common upper level for all Foundry ontologies. This means that BFO terms belong to each Foundry ontology, which conflicts with the idea that each term should belong to exactly one ontology. Terms not only from the BFO top-level ontology but also from mid- level ontologies such as IAO and OBI are re-used in multiple domain ontologies at lower levels. In like fashion, terms from more general domain ontologies such as the Cell Ontology are reused in ontologies such as the Cell Line Ontology and the Cell Cycle Ontology at lower levels. How, then, are do we save the principle that each term should belong to exactly one ontology? The answer is that orthogonality should be interpreted in such a way that each term should have its source – the original home where the term and its definition were manually crafted – in some one specific ontology, marked through the fact that the identifier for this ontology is included in the relevaxnt term IRI. This original term IRI should then travel with the term when it is reused in other ontologies, thereby preserving the link to the original home and at the same time networking the ontologies together, through reuse of terms not only in multiple ontology is-a hierarchies, but also in multiple term definitions.8 The hierarchy of ontologies reflects the hierarchy of sciences The way in which OBO Foundry ontologies are organized on successive levels represituation is analogous to what obtains in the organization of science. At any given stage of its development prior knowledge obtained through scientific research is organized hierarchically. Very general knowledge pertaining to general laws of physics or chemistry provides high-level starting points for scientific subdisciplines at successively lower levels of, for example, electrodynamics, quantum electrodynamics, organic chemistry, petroleum chemistry, neurochemistry, and so forth. The nodes in this (somewhat idealized) hierarchy then serve as starting points for huge numbers of further disciplines and subdisciplines, and also determine which bodies of prior knowledge are encapsulated in textbooks and which new results are reported in which journals. It is the titles of such textbooks and journals which tell us where we can find both prior knowledge and new results, just as it is the names of ontologies such as the Vaccine Ontology or the Infectious Disease Ontology which tell us where corresponding terms and definitions are to be found. 8 For example along the lines illustrated in https://bmcgenomics.biomedcentral.com/articles/10.1186/1471-2164-14- 513 Implications for deep learning All of this, perhaps surprisingly, has implications for contemporary artificial intelligence research. This is because one major shortfall of today’s dominant AI paradigm of autonomous deep learning is its inability to deal with prior knowledge. This is a problem, in areas such as biology and medicine, where huge amounts of prior knowledge already exist. It is a problem, too, for the public acceptance of the results of deep learning, since the use of neural networks creates what is in effect a black box, which means that we cannot understand how given results were achieved. This means, also, that we cannot assign responsibility – on the side of either the machine or its human programmers – when things go wrong. How, then, to supplement neural networks with prior knowledge in areas such as biology and medicine. To create representations inside the computer of entire disciplines, or of entire suites of disciplines and sub-disciplines, is of course an impractical goal.
Load more

Recommended publications
  • Original Article Text Mining in the Biocuration Workflow: Applications for Literature Curation at Wormbase, Dictybase and TAIR
    Database, Vol. 2012, Article ID bas040, doi:10.1093/database/bas040 ............................................................................................................................................................................................................................................................................................. Original article Text mining in the biocuration workflow: applications for literature curation at WormBase, dictyBase and TAIR Kimberly Van Auken1,*, Petra Fey2, Tanya Z. Berardini3, Robert Dodson2, Laurel Cooper4, Donghui Li3, Juancarlos Chan1, Yuling Li1, Siddhartha Basu2, Hans-Michael Muller1, Downloaded from Rex Chisholm2, Eva Huala3, Paul W. Sternberg1,5 and the WormBase Consortium 1Division of Biology, California Institute of Technology, 1200 E. California Boulevard, Pasadena, CA 91125, 2Northwestern University Biomedical Informatics Center and Center for Genetic Medicine, 420 E. Superior Street, Chicago, IL 60611, 3Department of Plant Biology, Carnegie Institution, 260 Panama Street, Stanford, CA 94305, 4Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331 and 5Howard Hughes Medical Institute, California Institute of Technology, 1200 E. California Boulevard, Pasadena, CA 91125, USA http://database.oxfordjournals.org/ *Corresponding author: Tel: +1 609 937 1635; Fax: +1 626 568 8012; Email: [email protected] Submitted 18 June 2012; Revised 30 September 2012; Accepted 2 October 2012 ............................................................................................................................................................................................................................................................................................
    [Show full text]
  • The Zebrafish Anatomy and Stage Ontologies: Representing the Anatomy and Development of Danio Rerio
    Van Slyke et al. Journal of Biomedical Semantics 2014, 5:12 http://www.jbiomedsem.com/content/5/1/12 JOURNAL OF BIOMEDICAL SEMANTICS RESEARCH Open Access The zebrafish anatomy and stage ontologies: representing the anatomy and development of Danio rerio Ceri E Van Slyke1*†, Yvonne M Bradford1*†, Monte Westerfield1,2 and Melissa A Haendel3 Abstract Background: The Zebrafish Anatomy Ontology (ZFA) is an OBO Foundry ontology that is used in conjunction with the Zebrafish Stage Ontology (ZFS) to describe the gross and cellular anatomy and development of the zebrafish, Danio rerio, from single cell zygote to adult. The zebrafish model organism database (ZFIN) uses the ZFA and ZFS to annotate phenotype and gene expression data from the primary literature and from contributed data sets. Results: The ZFA models anatomy and development with a subclass hierarchy, a partonomy, and a developmental hierarchy and with relationships to the ZFS that define the stages during which each anatomical entity exists. The ZFA and ZFS are developed utilizing OBO Foundry principles to ensure orthogonality, accessibility, and interoperability. The ZFA has 2860 classes representing a diversity of anatomical structures from different anatomical systems and from different stages of development. Conclusions: The ZFA describes zebrafish anatomy and development semantically for the purposes of annotating gene expression and anatomical phenotypes. The ontology and the data have been used by other resources to perform cross-species queries of gene expression and phenotype data, providing insights into genetic relationships, morphological evolution, and models of human disease. Background function, development, and evolution. ZFIN, the zebrafish Zebrafish (Danio rerio) share many anatomical and physio- model organism database [10] manually curates these dis- logical characteristics with other vertebrates, including parate data obtained from the literature or by direct data humans, and have emerged as a premiere organism to submission.
    [Show full text]
  • Biocuration 2016 - Posters
    Biocuration 2016 - Posters Source: http://www.sib.swiss/events/biocuration2016/posters 1 RAM: A standards-based database for extracting and analyzing disease-specified concepts from the multitude of biomedical resources Jinmeng Jia and Tieliu Shi Each year, millions of people around world suffer from the consequence of the misdiagnosis and ineffective treatment of various disease, especially those intractable diseases and rare diseases. Integration of various data related to human diseases help us not only for identifying drug targets, connecting genetic variations of phenotypes and understanding molecular pathways relevant to novel treatment, but also for coupling clinical care and biomedical researches. To this end, we built the Rare disease Annotation & Medicine (RAM) standards-based database which can provide reference to map and extract disease-specified information from multitude of biomedical resources such as free text articles in MEDLINE and Electronic Medical Records (EMRs). RAM integrates disease-specified concepts from ICD-9, ICD-10, SNOMED-CT and MeSH (http://www.nlm.nih.gov/mesh/MBrowser.html) extracted from the Unified Medical Language System (UMLS) based on the UMLS Concept Unique Identifiers for each Disease Term. We also integrated phenotypes from OMIM for each disease term, which link underlying mechanisms and clinical observation. Moreover, we used disease-manifestation (D-M) pairs from existing biomedical ontologies as prior knowledge to automatically recognize D-M-specific syntactic patterns from full text articles in MEDLINE. Considering that most of the record-based disease information in public databases are textual format, we extracted disease terms and their related biomedical descriptive phrases from Online Mendelian Inheritance in Man (OMIM), National Organization for Rare Disorders (NORD) and Orphanet using UMLS Thesaurus.
    [Show full text]
  • Human Disease Ontology 2018 Update: Classification, Content And
    Published online 8 November 2018 Nucleic Acids Research, 2019, Vol. 47, Database issue D955–D962 doi: 10.1093/nar/gky1032 Human Disease Ontology 2018 update: classification, content and workflow expansion Lynn M. Schriml 1,*, Elvira Mitraka2, James Munro1, Becky Tauber1, Mike Schor1, Lance Nickle1, Victor Felix1, Linda Jeng3, Cynthia Bearer3, Richard Lichenstein3, Katharine Bisordi3, Nicole Campion3, Brooke Hyman3, David Kurland4, Connor Patrick Oates5, Siobhan Kibbey3, Poorna Sreekumar3, Chris Le3, Michelle Giglio1 and Carol Greene3 1University of Maryland School of Medicine, Institute for Genome Sciences, Baltimore, MD, USA, 2Dalhousie University, Halifax, NS, Canada, 3University of Maryland School of Medicine, Baltimore, MD, USA, 4New York University Langone Medical Center, Department of Neurosurgery, New York, NY, USA and 5Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA Received September 13, 2018; Revised October 04, 2018; Editorial Decision October 14, 2018; Accepted October 22, 2018 ABSTRACT INTRODUCTION The Human Disease Ontology (DO) (http://www. The rapid growth of biomedical and clinical research in re- disease-ontology.org), database has undergone sig- cent decades has begun to reveal novel cellular, molecular nificant expansion in the past three years. The DO and environmental determinants of disease (1–4). However, disease classification includes specific formal se- the opportunities for discovery and the transcendence of mantic rules to express meaningful disease mod- knowledge between research groups can only be realized in conjunction with the development of rigorous, standard- els and has expanded from a single asserted clas- ized bioinformatics tools. These tools should be capable of sification to include multiple-inferred mechanistic addressing specific biomedical data nomenclature and stan- disease classifications, thus providing novel per- dardization challenges posed by the vast variety of biomed- spectives on related diseases.
    [Show full text]
  • Biocuration - Mapping Resources and Needs [Version 2; Peer Review: 2 Approved]
    F1000Research 2020, 9(ELIXIR):1094 Last updated: 22 JUL 2021 RESEARCH ARTICLE Biocuration - mapping resources and needs [version 2; peer review: 2 approved] Alexandra Holinski 1, Melissa L. Burke 1, Sarah L. Morgan 1, Peter McQuilton 2, Patricia M. Palagi 3 1European Molecular Biology Laboratory, European Bioinformatics Institute (EMBL-EBI), Hinxton, Cambridgeshire, CB10 1SD, UK 2Oxford e-Research Centre, Department of Engineering Science, University of Oxford, Oxford, Oxfordshire, OX1 3QG, UK 3SIB Swiss Institute of Bioinformatics, Lausanne, 1015, Switzerland v2 First published: 04 Sep 2020, 9(ELIXIR):1094 Open Peer Review https://doi.org/10.12688/f1000research.25413.1 Latest published: 02 Dec 2020, 9(ELIXIR):1094 https://doi.org/10.12688/f1000research.25413.2 Reviewer Status Invited Reviewers Abstract Background: Biocuration involves a variety of teams and individuals 1 2 across the globe. However, they may not self-identify as biocurators, as they may be unaware of biocuration as a career path or because version 2 biocuration is only part of their role. The lack of a clear, up-to-date (revision) report profile of biocuration creates challenges for organisations like ELIXIR, 02 Dec 2020 the ISB and GOBLET to systematically support biocurators and for biocurators themselves to develop their own careers. Therefore, the version 1 ELIXIR Training Platform launched an Implementation Study in order 04 Sep 2020 report report to i) identify communities of biocurators, ii) map the type of curation work being done, iii) assess biocuration training, and iv) draw a picture of biocuration career development. 1. Tanya Berardini , Phoenix Bioinformatics, Methods: To achieve the goals of the study, we carried out a global Fremont, USA survey on the nature of biocuration work, the tools and resources that are used, training that has been received and additional training 2.
    [Show full text]
  • Improving the Gene Ontology Resource to Facilitate More Informative Analysis and Interpretation of Alzheimer’S Disease Data
    G C A T T A C G G C A T genes Article Improving the Gene Ontology Resource to Facilitate More Informative Analysis and Interpretation of Alzheimer’s Disease Data Barbara Kramarz 1 , Paola Roncaglia 2 , Birgit H. M. Meldal 2 , Rachael P. Huntley 1 , Maria J. Martin 2, Sandra Orchard 2, Helen Parkinson 2, David Brough 3, Rina Bandopadhyay 4, Nigel M. Hooper 3 and Ruth C. Lovering 1,* 1 UCL Institute of Cardiovascular Science, University College London, Rayne Building, 5 University Street, London WC1E 6JF, UK; [email protected] (B.K.); [email protected] (R.P.H.) 2 European Bioinformatics Institute (EMBL-EBI), European Molecular Biology Laboratory, Wellcome Genome Campus, Hinxton, Cambridge CB10 1SD, UK; [email protected] (P.R.); [email protected] (B.H.M.M.); [email protected] (M.J.M.); [email protected] (S.O.); [email protected] (H.P.) 3 Division of Neuroscience and Experimental Psychology, School of Biological Sciences, Faculty of Biology, Medicine and Health, Manchester Academic Health Science Centre, University of Manchester, AV Hill Building, Oxford Road, Manchester M13 9PT, UK; [email protected] (D.B.); [email protected] (N.M.H.) 4 UCL Queen Square Institute of Neurology and Reta Lila Weston Institute of Neurological Studies, 1 Wakefield Street, London WC1N 1PJ, UK; [email protected] * Correspondence: [email protected] or [email protected]; Tel.: +44-207-679-6965 Received: 31 October 2018; Accepted: 23 November 2018; Published: 29 November 2018 Abstract: The analysis and interpretation of high-throughput datasets relies on access to high-quality bioinformatics resources, as well as processing pipelines and analysis tools.
    [Show full text]
  • Biocuration Experts on the Impact of Duplication and Other Data Quality Issues in Biological Databases
    Genomics Proteomics Bioinformatics 18 (2020) 91–103 Genomics Proteomics Bioinformatics www.elsevier.com/locate/gpb www.sciencedirect.com PERSPECTIVE Quality Matters: Biocuration Experts on the Impact of Duplication and Other Data Quality Issues in Biological Databases Qingyu Chen 1,*, Ramona Britto 2, Ivan Erill 3, Constance J. Jeffery 4, Arthur Liberzon 5, Michele Magrane 2, Jun-ichi Onami 6,7, Marc Robinson-Rechavi 8,9, Jana Sponarova 10, Justin Zobel 1,*, Karin Verspoor 1,* 1 School of Computing and Information Systems, University of Melbourne, Melbourne, VIC 3010, Australia 2 European Molecular Biology Laboratory, European Bioinformatics Institute (EMBL-EBI), Wellcome Trust Genome Campus, Cambridge CB10 1SD, UK 3 Department of Biological Sciences, University of Maryland, Baltimore, MD 21250, USA 4 Department of Biological Sciences, University of Illinois at Chicago, Chicago, IL 60607, USA 5 Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA 6 Japan Science and Technology Agency, National Bioscience Database Center, Tokyo 102-8666, Japan 7 National Institute of Health Sciences, Tokyo 158-8501, Japan 8 Swiss Institute of Bioinformatics, CH-1015 Lausanne, Switzerland 9 Department of Ecology and Evolution, University of Lausanne, CH-1015 Lausanne, Switzerland 10 Nebion AG, 8048 Zurich, Switzerland Received 8 December 2017; revised 24 October 2018; accepted 14 December 2018 Available online 9 July 2020 Handled by Zhang Zhang Introduction assembled, annotated, and ultimately submitted to primary nucleotide databases such as GenBank [2], European Nucleo- tide Archive (ENA) [3], and DNA Data Bank of Japan Biological databases represent an extraordinary collective vol- (DDBJ) [4] (collectively known as the International Nucleotide ume of work.
    [Show full text]
  • The Evaluation of Ontologies: Editorial Review Vs
    The Evaluation of Ontologies: Editorial Review vs. Democratic Ranking Barry Smith Department of Philosophy, Center of Excellence in Bioinformatics and Life Sciences, and National Center for Biomedical Ontology University at Buffalo from Proceedings of InterOntology 2008 (Tokyo, Japan, 26-27 February 2008), 127-138. ABSTRACT. Increasingly, the high throughput technologies used by biomedical researchers are bringing about a situation in which large bodies of data are being described using controlled structured vocabularies—also known as ontologies—in order to support the integration and analysis of this data. Annotation of data by means of ontologies is already contributing in significant ways to the cumulation of scientific knowledge and, prospectively, to the applicability of cross-domain algorithmic reasoning in support of scientific advance. This very success, however, has led to a proliferation of ontologies of varying scope and quality. We define one strategy for achieving quality assurance of ontologies—a plan of action already adopted by a large community of collaborating ontologists—which consists in subjecting ontologies to a process of peer review analogous to that which is applied to scientific journal articles. 1 From OBO to the OBO Foundry Our topic here is the use of ontologies to support scientific research, especially in the domains of biology and biomedicine. In 2001, Open Biomedical Ontologies (OBO) was created by Michael Ashburner and Suzanna Lewis as an umbrella body for the developers of such ontologies in the domain of the life sciences, applying the key principles underlying the success of the Gene Ontology (GO) [GOC 2006], namely, that ontologies be (a) open, (b) orthogonal, (c) instantiated in a well-specified syntax, and such as (d) to share a common space of identifiers [Ashburner et al.
    [Show full text]
  • ISB Newsletter
    ISB June 2012 . INTERNATIONAL SOCIETY ISB FOR BIOCURATION June, 2012 In this issue: - Save the Date: ISB2013 Save the Date - B3CB Workshop Report - ISB Membership Renewal for ISB2013! - 2012 Elections: ISB Executive Committee The 6th International Biocuration Conference - Jack Leunissen will be held on April 7-10, - Latest Advances from ISB Research Teams 2013, at Churchill College, - Job Opportunities Cambridge, UK. Training Activities: B3CB Workshop Report B3CB was a workshop aimed at developing common resources and infrastructures for training of experts as well as users of bioinformatics tools and resources. It was held in Uppsala (Sweden), and chaired by Terri Attwood (Manchester University, ISB Executive Committee member). Pascale Gaudet introduced the ISB. We believe that this initiative will help the ISB provide more support for training, a part of the our mission statement. The workshop brought together members of the executive committees of several ISB Membership Renewal organizations in Bioinformatics, Biotechnology, Biocuration and Computational Biology (B3CB): For many of us, membership fees are due - EMBnet (Global Bioinformatics Network) – Terri Attwood this month! Notices for renewal of ISB - ISCB (International Society for Computational Biology) – Reinhard Schneider Memberships will be reaching inboxes in - APBioNet (Asia-Pacific Bioinformatics Network) – Christian Schönbach the coming days. Your ISB Membership - ASBCB (African Society for Bioinformatics & Computational Biology) – Nicky Mulder goes to support activities such as the - SoIBio (Iberoamerican Society for Bioinformatics) – Oswaldo Trelles International Biocuration Conference - ISB (International Society for Biocuration) – Pascale Gaudet (travel support was provided for 10 - EBI/ELIXIR (European Bioinformatics Institute) – Cath Brooksbank attendees to ISB2012), BioDBCore (in - NBIC (Netherlands Bioinformatics Centre) – Celia van Gelder collaboration with BioSharing), and to - SeqAhead (Next-Gen Sequencing Data Analysis Network) – Erik Bongcam-Rudloff establish links with other groups.
    [Show full text]
  • Applications of an Ontology Engineering Methodology Accessing Linked Data for Dialogue-Based Medical Image Retrieval
    Applications of an Ontology Engineering Methodology Accessing Linked Data for Dialogue-Based Medical Image Retrieval Daniel Sonntag, Pinar Wennerberg, Sonja Zillner DFKI - German Research Center for AI, Siemens AG [email protected] [email protected] [email protected] Abstract of medication and treatment plan is foreseen. This requires This paper examines first ideas on the applicability of the medical images to be annotated accordingly so that the Linked Data, in particular a subset of the Linked Open Drug radiologists can obtain all the necessary information Data (LODD), to connect radiology, human anatomy, and starting with the image. Moreover, the image information drug information for improved medical image annotation needs to be interlinked in a coherent way to enable the and subsequent search. One outcome of our ontology radiologist to trace the dependencies between observations. engineering methodology is the semi-automatic alignment Linked Data has the potential to provide easier access to between radiology-related OWL ontologies (FMA and significantly growing, publicly available related data sets, RadLex). These can be used to provide new connections in such as those in the healthcare domain: in combination the medicine-related linked data cloud. A use case scenario with ontology matching and speech dialogue (as AI is provided that involves a full-fletched speech dialogue system as AI application and additionally demonstrates the systems) this should allow medical experts to explore the benefits of the approach by enabling the radiologist to query interrelated data more conveniently and efficiently without and explore related data, e.g., medical images and drugs. having to switch between many different applications, data The diagnosis is on a special type of cancer (lymphoma).
    [Show full text]
  • Modularization for the Cell Ontology
    Modularization for the Cell Ontology Chris Mungall, Melissa Haendel, Amelia Ireland, Shahid Manzoor, Terry Meehan, David Osumi-Sutherland, Carlo Torniai, Alexander Diehl Outline • The Cell Ontology and its neighbors • Handling multiple species • Extracting modules for taxonomic contexts • The Oort – OBO Ontology Release Tool Inter-ontology axioms Source Ref Axiom cl go ‘mature eosinophil’ SubClassOf capable of some ‘respiratory burst’ go cl ‘eosinophil differentiation’ EquivalentTo differentiation and has_target some eosinophil * cl pro ‘Gr1-high classical monocyte’ EquivalentTo ‘classical monocyte’ and has_plasma_membrane_part some ‘integrin-alpha-M’ and …. cl chebi melanophage EquivalentTo ‘tissue-resident macrophage’ and has_part some melanin and … cl uberon ‘epithelial cell of lung’ EquivalentTo ‘epithelial cell’ and part_of some lung uberon cl muscle tissue EquivalentTo tissue and has_part some muscle cell clo cl HeLa cell SubClassOf derives_from some (‘epithelial cell’ and part_of some cervix) fbbt cl fbbt:neuron EquivalentTo cl:neuron and part_of some Drosophila ** * experimental extension ** taxon-context bridge axiom The cell ontology covers multiple taxonomic contexts Ontology Context # of links to cl MA adult mouse 1 FMA adult human 658 XAO frog 63 AAO amphibian ZFA zebrafish 391 TAO bony fish 385 FBbt fruitfly 53 PO plants * CL all cells - Xref macros: an easy way to formally connect to broader taxa treat-xrefs-as-genus-differentia: CL part_of NCBITaxon:7955! …! [Term]! id: ZFA:0009255! name: amacrine cell! xref: CL:0000561!
    [Show full text]
  • Investigating Term Reuse and Overlap in Biomedical Ontologies Maulik R
    Investigating Term Reuse and Overlap in Biomedical Ontologies Maulik R. Kamdar, Tania Tudorache, and Mark A. Musen Stanford Center for Biomedical Informatics Research, Department of Medicine, Stanford University ABSTRACT CT, to spare the effort in creating already existing quality We investigate the current extent of term reuse and overlap content (e.g., the anatomy taxonomy), to increase their among biomedical ontologies. We use the corpus of biomedical interoperability, and to support its use in electronic health ontologies stored in the BioPortal repository, and analyze three types records (Tudorache et al., 2010). Another benefit of the of reuse constructs: (a) explicit term reuse, (b) xref reuse, and (c) ontology term reuse is that it enables federated search Concept Unique Identifier (CUI) reuse. While there is a term label engines to query multiple, heterogeneous knowledge sources, similarity of approximately 14.4% of the total terms, we observed structured using these ontologies, and eliminates the need for that most ontologies reuse considerably fewer than 5% of their extensive ontology alignment efforts (Kamdar et al., 2014). terms from a concise set of a few core ontologies. We developed For the purpose of this work, we define as term reuse an interactive visualization to explore reuse dependencies among the situation in which the same term is present in two or biomedical ontologies. Moreover, we identified a set of patterns that more ontologies either by direct use of the same identifier, indicate ontology developers did intend to reuse terms from other or via explicit references and mappings. We define as term ontologies, but they were using different and sometimes incorrect overlap the situation in which the term labels in two or more representations.
    [Show full text]